Tweets Become Harder to Believe as Labels Change Meaning

In the 24 hours after Twitter last week eliminated the blue check mark that historically served as a means of identifying public agencies, at least 11 new accounts began impersonating the Los Angeles Police Department.

More than 20 purported to be various agencies of the federal government. Someone pretending to be the mayor of New York City promised to create a Department of Traffic and Parking Enforcement and slash police funding by 70 percent.

Elon Musk’s decision to stop giving check marks to people and groups verified to be who they said were, and instead offering them to anyone who paid for one, is the latest tumult at Twitter, the social media giant he has vowed to remake since he acquired it last year for $44 billion.

The changes have convulsed a platform that once seemed indispensable for following news as it broke around the world. The information on Twitter is now increasingly unreliable. Accounts that impersonate public officials, government agencies and celebrities have proliferated. So have propaganda and disinformation that threaten to further erode trust in public institutions. The consequences are only beginning to emerge.

Alyssa Kann, a research associate at the Atlantic Council’s Digital Forensic Research Lab, said Twitter under Mr. Musk was systematically dismantling safeguards that had been put in place over years of consideration and controversy.

“When there are so many things going wrong at once, it’s like: Which fire do you put out first?” she said.

After a public dispute with NPR, which Twitter falsely labeled state-affiliated media, the platform last week removed all labels that had identified state-owned media, including those controlled by authoritarian states like Russia, China and Iran.

That, coupled with a decision to stop blocking recommendations for them, has coincided with a spike in engagement for many of these accounts, according to research by the Digital Forensic Research Lab and another organization that studies disinformation, Reset, which is based in London.

In Sudan, new accounts on Twitter are falsely representing both sides of the civil war that has erupted there. One account that, presumably, bought a blue check mark falsely proclaimed the death of Lt. Gen. Mohamed Hamdan, the leader of the rebel Rapid Support Forces. More than 1.7 million people viewed the tweet.

Twitter’s new head of trust and safety, Ella Irwin, did not respond to a request for comment on the changes and their consequences.

Twitter has always been a font of misinformation and worse, but the previous policies sought to inform readers of the sources of content and limit the most egregious instances. The debut of verified accounts at Twitter in 2009 is usually associated with Tony La Russa, a major-league baseball manager who sued Twitter for trademark infringement and other claims after being impersonated on the platform.

Over time, verified accounts with blue check marks steered users to official sources and real people. Labeling news organizations as state media indicated that the accounts reflected a certain point of view.

Impersonators became a problem almost immediately after Mr. Musk took the helm in November and offered to sell the check marks to anyone who subscribed for the monthly fee. He backtracked after companies like Eli Lilly and PepsiCo grappled with seemingly verified spoof accounts promising free insulin and praising the superiority of Coca-Cola.

By last week, Twitter had begun removing the blue check marks from companies, government agencies, news organizations and others who did not agree to pay. It appears that many chose not to sign up, though Twitter has not disclosed any figures.

Some cheered the changes.

“Now you can even find me in the search,” tweeted Margarita Simonyan, the editor in chief of RT, the Russian state television network that has been accused of rampant misinformation and hate speech aimed at Ukraine. She signed off the tweet by saying, “Brotherly, Elon @elonmusk, from the heart.”

Twitter’s algorithms previously excluded accounts labeled state officials or media from recommendations, dampening engagement. According to Reset, 124 accounts belonging to Russian state media have received on average 33 percent more exposure in views and impressions after the changes, which took effect in late March.

They include accounts like that of Dmitri A. Medvedev, the former president of Russia and deputy chairman of the country’s security council, who posted a distorted photograph of President Biden on Tuesday, calling him in English “a daring geezer.”

When an account argued this month that Twitter was amplifying Russia’s genocidal propaganda toward Ukraine, Mr. Musk replied dismissively: “All news is to some degree propaganda. Let people decide for themselves.” (The account he was responding to has since been suspended.)

Researchers said the abrupt changes in how the check marks are obtained threatened, at a minimum, to create confusion. They could also undermine trust in a tool for communication during crises like natural disasters.

The main account of the Los Angeles Police Department has a gray check mark, which Twitter created for “legacy accounts,” but not all of its various bureaus do — the Hollywood division, for example. In addition to providing blue check marks for $8 a month, Twitter has invited organizations to pay $1,000 to receive gold marks for multiple accounts. For a time, at least, one was extended to a Disney Junior impostor account that tweeted racist and vulgar language.

“This is going to be chaos for emergency services,” tweeted Marc-André Argentino, a research fellow at the London-based International Center for the Study of Radicalization.

Mr. Argentino tracked examples showing an account impersonating the mayor of Chicago replying to one impersonating the city’s Department of Transportation. Another had New York City’s actual government-run account arguing with an impostor.

“Yes this is funny, let us all laugh,” Mr. Argentino wrote. “Now take two seconds and go back to any mass casualty incident in a major city, or a natural disaster, or any crisis/critical incident when people turn to official sources of information in times need & think of the harm that this can cause.”

On Friday, the comedian George Carlin’s daughter, Kelly Carlin, tweeted an accusation that someone was masquerading as the account she runs for her late father, even using the same profile photo and claiming to be her.

“HERE IT BEGINS,” she wrote, later complaining after several unsuccessful attempts to have the impostor account removed that “Twitter is broken.” The spoof account was still up on Wednesday, with nine followers.

Josh Boerman, who co-hosts a pop culture podcast, “The Worst of All Possible Worlds,” was the source of the account impersonating Mayor Eric Adams of New York, promising to create a traffic and parking department and cut police funding.

Mr. Boerman said he had tried hard to leave obvious hints that he was an impersonator. His tweet thread included unrealistic scenarios where all police officers’ guns were melted down and sold for scrap, with the proceeds going to the parks department. He made up an organization with a ludicrous name: the New York City Porcine Benevolent Association. He promoted his podcast to his relatively small Twitter following of 1,700 users.

“Pretty much everybody got that it was a joke immediately, which was my hope — I wasn’t trying to mislead anyone,” Mr. Boerman said. “The point was that this can be both a joke on the state of the network right now as well as an opportunity to think about the way that media is disseminated and how we think about our public figures.”

The removal of the blue verification badges caused “immediate and pure chaos,” but the novelty eventually wore off, he said. His profile name is now “bosh (not mayor anymore).” He said he was careful to confirm any announcement he saw on Twitter using other sources.

“The problem comes when you have accounts that maybe have hundreds of thousands of followers and are positioning themselves as the real thing,” Mr. Boerman said. “Twitter’s approach of ‘Well, if people pay for verification, certainly they must be legit’ is so inane I don’t even know how to put words to it.”